Patterns for Fault Tolerant Software

Conclusion

The previous five chapters contain patterns that can help make your software designs more fault tolerant. Chapter 4 contained architectural principles that you should use to structure your overall solution. Chapter 5 presented patterns to help detect the presence of errors and faults before failures occur. Chapters 6 and 7 gave ways to process errors. Sometimes the processing results in the system execution state changing, and sometimes the errors and faults can be processed in a way that doesn't change the execution flow. Chapter 8 contained patterns to help treat the faults and return the erroneous component to availability. This chapter shows the complete pattern language and gives an example of how the patterns can work together.

A Pattern Language for Fault Tolerant Software

Each chapter began with a description of the patterns and how they related to each other and how they worked together to make a system more fault tolerant. Linkages between the patterns in different chapters were not highlighted, except in the individual pattern descriptions. See back of book for a map of all of the patterns. Within this overall pattern map you can see some of the linkages between the patterns of the individual phases.

This figure represents only one possible combination of these patterns into a usable language. Others are possible, because the choice of patterns to apply to a given design problem depends on the context and the nature of the system, as illustrated through the example ...